Nobel Lecture: From spin waves to giant magnetoresistance and beyond

Abstract

The “Institute for Magnetism” within the department for solid-state physics at the research center in Julich, Germany, which I joined in 1972, was founded in 1971 by Professor W. Zinn. The main research topic was the exploration of the model magnetic semiconductors EuO and EuS with Curie temperatures Tc=60 K and Tc =17 K, respectively. As I had been working with light scattering LS techniques before I came to Julich, I was very much interested in the observation of spin waves in magnetic materials by means of LS. LS can be performed with grating spectrometers, which is called Raman spectroscopy, and alternatively by Brillouin light scattering BLS spectroscopy. In the latter case, a Fabry-Perot FP interferometer is used for the frequency analysis of the scattered light see righthand side of Fig. 1 . The central part consists of two FP mirrors whose distance is scanned during operation. BLS spectroscopy is used when the frequency shift of the scattered light is small below 100 GHz , as expected for spin waves in ferromagnets. In the early 1970s, an interesting instrumental development took place in BLS, namely, the invention of the multipass operation, and later, the combination of two multipass interferometers in tandem. The inventor was Dr. J. A. Sandercock in Zurich. Since we had the opportunity to install a new laboratory, we decided in favor of BLS, initially using a single three-pass instrument as displayed on the righthand side of Fig. 1. With this, we started investigating spin waves in EuO. We indeed were able to find and identify the expected spin waves as shown by the peaks in Fig. 1 marked green . Different intensities on the Stokes S and antiStokes aS side were known from other work to be due to the magneto-optic interaction of light with the spin waves. The peaks marked red remained a puzzle for some time until good luck came to help us. Good luck in this case was a breakdown of the system, a repair and unintentional interchange of the leads when reconnecting the magnet to the power supply. To our surprise S and aS side were now reversed. To understand what this means, one has to know that classically S and aS scattering is related to the propagation direction of the observed mode, which is opposite for the two cases. This can be understood from the corresponding Doppler shift, which is to higher frequencies when the wave travels towards the observer and down when away from him. The position of the observer here would be the same as of the viewer in Fig. 1. The appearance of the red peak in the spectra on only either the S or the aS side can be explained by an unidirectional propagation of the corresponding spin wave along the surface of the sample. It can be reversed by reversing B0 and M. The unidirectional behavior of the wave can be understood on the basis of symmetry. For this, one has to know that axial vectors which appear in nature, such as B and M on the left-hand side of Fig. 1, reverse their sign under time inversion and so does the sense of the propagation of the surface wave as indicated. The upper and lower parts of Fig. 1, on the left-hand side therefore are linked by time inversion symmetry, which is valid without damping. Hence, the unidirectional behavior reflects the symmetry of the underlying system. Finally, the observed wave could be identified as the DamonEshbach DE surface mode known from theory and from microwave experiments. From the magnetic parameters of EuO, one predicts in the present case that the penetration depth of the DE mode will be a few 100 A. Sample thickness d is of the order of mm. Therefore, for the present purpose, EuO is opaque. In this case, the wave traveling on the backside of the sample in the opposite direction to the wave on the front side cannot be seen in this experiment. BLS is then either S or aS but not both at the same time. Due to all of these unique features, the results of Fig. 1 have also been chosen as examples for current research in magnetism in a textbook on “Solid State Physics” Ibach and Luth, 1995, p. 186 .